Question

我是avro的新手，在尝试使用avro别名属性时，我遇到了错误。

查询：select department_id , office_name from test.depart_alias;

SemanticException [错误10004]：行1:23无效的表别名或列引用'office_name':(可能的列名称为：department_id，department_name）

我的avsc格式的json文件是

{
  "type" : "record",
  "name" : "departments",
  "doc" : "Sqoop import of departments",
  "fields" : [ {
    "name" : "department_id",
    "type" : [ "null", "int" ],
    "default" : null,
    "columnName" : "department_id",
    "sqlType" : "4"
  }, {
    "name" : "department_name",
    "type" : [ "null", "string" ],
    "default" : null,
    "aliases" : [ "office_name" ],
    "columnName" : "department_name",
    "sqlType" : "12"
  } ],
  "tableName" : "departments"

我想使用office name作为department_name的别名。我应该如何实施呢？

由于

Answer 1

如果您的旧列字段名为department_name，则您的架构应如下所示

{
  "type" : "record",
  "name" : "departments",
  "doc" : "Sqoop import of departments",
  "fields" : [ {
    "name" : "department_id",
    "type" : [ "null", "int" ],
    "default" : null,
    "columnName" : "department_id",
    "sqlType" : "4"
  }, {
    "name" : "office_name",
    "type" : [ "null", "string" ],
    "default" : null,
    "aliases" : [ "department_name" ], <- old columns here
    "columnName" : "department_name",
    "sqlType" : "12"
  } ],
  "tableName" : "departments"

旧列名或“别名”应该在别名数组中。

如何实现avro别名

1 个答案: